Ane Bærent Fisker
|Fisker, Ane Bærent. 2011. Evaluation of the WHO vitamin A supplementation policy in Guinea-Bissau. Bandim Health Project, Statens Serum Institut. Department of Biostatistics, Institute of Public Health, Aarhus University. 179 p.
|Vitamin A supplementation (VAS) is estimated to reduce overall mortality by 24% when provided to children aged 6-59 months at 4-6-monthly intervals in low income countries. WHO recommends integration of VAS with immunisation services in areas where vitamin A deficiency is a public health problem. The effect of this policy on overall mortality has not been evaluated. Observational data and trials powered to study other outcomes than mortality, suggest that VAS with the inactivated diphtheria-tetanus-pertussis (DTP) vaccine may be associated with higher mortality due to a negative interaction between VAS and DTP vaccines in girls.
The thesis consists of 4 papers exploring interactions between vitamin A and vaccines and whether he effects differ by sex. The population studied is children in the study area of the Bandim Health Project (BHP) in Guinea-Bissau.
Paper I evaluates the strategy of providing VAS in biannual campaigns. The study is based on observational data since the strategy is already implemented in Guinea-Bissau and hence it would not be ethically acceptable to conduct a randomised trial. The paper describes mortality in children 6-35 months according to participation in two vitamin A campaigns. Both campaigns had poor overage (58% and 68%) in the followed age groups. Mortality was similar in supplemented and non-supplemented children, the adjusted mortality rate ratio (MRR) being 0.82 (0.48-1.40). We tested the hypothesis that the effect of vitamin A varies depending on the last vaccine at the time of the campaign. As hypothesised we found that the MRRs differed with East vaccine: among children who had received a measles vaccine as their last vaccine prior to the campaign, supplemented children had lower mortality than non-supplemented children, adjusted (O. 1.4-0.85). Among children who had received DTP as their last vaccine prior to the campaign, supplemented children had no survival advantage MRR=1.29 (0.51-3.21), p=0.04 for different effect of VAS in children who had received MV and DTP.
Papers II-IV are based on the first randomised trial of the strategy of providing VAS at vaccination contacts. The trial was conducted from 2007-2011 in the urban and rural study areas of the BHP. We examined the effect of VAS with vaccines on mortality, adverse reactions and growth. Children aged 6-23 (6-17 months in the urban area) were randomized to VAS or placebo at vaccination contacts, and followed during the subsequent year or until first visit after 25 February 2011. A total of 7585 children were enrolled.
Adverse reactions were studied in a subgroup of 1673 children enrolled in the urban area. We identified no severe ad verse reactions, but found that VAS had sex-differential effects on increased intracranial pressure (a common adverse reaction to V AS), and sex- and vaccine-differential effects on adverse reactions to vaccines. The results support that the effect of VAS on the immune system vary depending on sex and vaccine type.
In spite of a high prevalence of vitamin A deficiency VAS had no overall effect on mortality (MRR=1.02 (0.69;1.51)). The effect of VAS differed by sex: VAS tended to increase mortality in boys (MRR=1.71(0.96;3.06)) and lower mortality in girls (MRR=0.62 (0.35;1.09) (p=0.01 for interaction between VAS and sex). The effect of VAS also varied significantly with season, previous VAS and urban-rural enrolment: VAS was associated with a beneficial effect in girls in the dry season and in girls who had received VAS on a previous occasion, but a detrimental effect in boys in the rural area.
In a subgroup of 1596 children enrolled in the urban area we studied VAS affected growth and whether the effect differed after vaccination with a live compared with an inactivated vaccine overall and separately for boys and girls. VAS length-for-age overall, but had no effect on weight-for-age or arm-circumference-for-age. In 443 children who had received an inactivated vaccine at enrolment. VAS benefitted the growth of boys during the first 6 months whereas girls tended to have a negative effect. In contrast, in 1001 children who had received 21 live vaccine VAS tended to benefit the growth of girls but not boys. Thus the effect of VAS differed by sex and vaccine, resulting in significant three-way interactions for weight-for-age (p=0.008) and arm-circumference-for-age (p=0.005), but not for length-for-age (p=0.08).
The policy is based on the expectation of an overall reduction in mortality. We found none in Guinea-Bissau, but identified several subgroups in which the effect of VAS on mortality is not beneficial, underlining that the current recommendation of VAS to all children after 6 months of age and at all vaccination contacts may not be the best way to reduce child mortality and may increase mortality in subgroups. The heterogeneity of the effect of VAS was also reflected in adverse reactions and growth.